Regular Article Molecular Characterization of Two Monkey Dihydrodiol Dehydrogenases

Drug Metabol. Pharmacokin. 17 (4): 348–356 (2002).

Yu HIGAKI1,TakeshiKAMIYA1,NoriyukiUSAMI1,SyunichiSHINTANI1,HiroakiSHIRAISHI2, Shuhei ISHIKURA1,IkuoYAMAMOTO3 and Akira HARA1 1Laboratory of Biochemistry, Gifu Pharmaceutical University, Gifu, Japan 2Department of Pharmacology, Fujita Health University School of Medicine, Aichi, Japan 3Faculty of Health Science, Kyushu University of Health and Welfare, Nobeoka, Japan

Summary: Japanese monkey liver contains multiple forms of dihydrodiol dehydrogenase with 3(20)a- hydroxysteroid dehydrogenase activity. Here we have puriˆed the major and minor forms (DD1 and DD4) of the enzyme from Cynomolgus monkey liver, and isolated cDNA species for the two enzyme forms by reverse transcription-PCR. The cDNAs encoded proteins comprising of 323 amino acids, in which the sequence identity between DD1 and DD4 was 83z. The sequences deduced from the cDNAs for DD1 and DD4 perfectly matched the partial sequences of peptides derived from the respective enzymes. We also isolated the cDNAs for DD1 and DD4 of Japanese monkey liver, which had almost identical amino acid sequences with those of the respective enzymes of Cynomolgus monkey liver. The monkey DD1s and DD4s showed the highest sequence identity (94z)withAKR1C1andAKR1C4,respectively, of four isoenzymes of human 3(20)a-hydroxysteroid dehydrogenase, which belongs to the aldo-keto reductase family. The substrate speciˆcity and inhibitor sensitivity of the puriˆed recombinant Cynomolgu monkey DD1 and Japanese monkey DD4 were also essentially identical to those of the recombinant AKR1C1 and AKR1C4, respectively, indicating that DD1 and DD4 are homologues of human AKR1C1 and AKR1C4, respectively. The mRNA for DD1 was detected only in liver, kidney, intestine and adrenal gland among Japanese monkey tissues, and that for DD4 was expressed in liver and kidney. These tissue distribution patterns diŠer from those of human AKR1C1 and AKR1C4, which are expressed ubiquitously and liver-speciˆc, respectively. In addition, no mRNA for an enzyme corresponding to another isoenzyme (AKR1C2) of the human enzyme was detected in livers of the two monkey strains. The results suggest a diŠerence in the metabolism of steroids and xenobiotics mediated by 3(20)a- hydroxysteroid dehydrogenase isoenzymes between monkeys and humans.

Key words: dihydrodiol dehydrogenase; indanol dehydrogenase; 3a-hydroxysteroid dehydrogenase; 20a- hydroxysteroid dehydrogenase; aldo-keto reductase family

whereas more than three DDs have been puriˆed from Introduction other species. The major form of hepatic DD is identical Dihydrodiol dehydrogenase (DD, EC 1.3.1.20) with 3a-HSDs in rat3,4) and hamster,5) but with 17b- catalyzes the NADP+-linked oxidation of trans-di- HSDs in mouse6) and guinea pig,7) and NADP+-depen- hydrodiols of aromatic hydrocarbons to the cor- dent D-xylose dehydrogenase in pig.9) responding catechols, and has been shown to be in- In humans, the major form of hepatic DD is 3a- volved in metabolic activation of the hydrocarbons HSD,10,11) but it exists as four isoenzymes, which exhibit through auto-oxidation of the catechol products.1,2) In sequence identity of more than 83z andbelongtothe mammalian tissues the enzyme exists in multiple forms, aldo-keto reductase (AKR) superfamily.12) According to which have been shown to be identical with 3a-, 17b- the nomenclature for the AKR family, these isoenzymes and 20a-hydroxysteroid dehydrogenases (HSDs), alde- are named AKR1C1, AKR1C2, AKR1C3 and AKR1C4, hyde reductase andWor aldose reductase.3–9) However, which correspond to previously known 3(20)a-HSD,13) there are species diŠerences in the number of multiple 3a-HSD type 3,14) 3a-HSD type 2,15,16) and 3a-HSD type forms of hepatic DD and in the nature of its major 1,15,17) respectively. These isoenzymes diŠer in their a‹n- form. For example, only two DDs are detected in rat, ity for steroidal substrates, inhibitor sensitivity and

Received; June 21, 2002, Accepted; August 22, 2002 To whom correspondence should be addressed: Prof. Akira HARA,Ph.D.,Laboratory of Biochemistry, Gifu Pharmaceutical University, Mitahora-higashi, Gifu 502-8585, Japan. Tel. & Fax. +81-58-237-8586, E-mail: hara＠gifu-pu.ac.jp

348 Monkey Dihydrodiol Dehydrogenases 349 eŠects of activators, but exhibit bifunctional activities monkey liver,25) and designated as cyDD1 and cyDD4, of 17b-and20a-HSDs.18) AKR1C1 is also called 20a- respectively. The enzymes were digested with LysC or HSD because of its high 20a-HSD activity,19) and cleaved by BrCN after reductive S-alkylation, and the AKR1C3 has been demonstrated to be identical with peptides isolated on reverse phase HPLC were se- 17b-HSD type 520) and prostaglandin F synthase.21,22) In quenced by automated Edman degradation with a 473A addition, all the isoenzymes, except AKR1C3, reduce protein sequencer (Applied Biosystems) as described ketone-containing drugs, although the substrate previously.17) speciˆcity for drugs diŠers among the isoenzymes.23,24) cDNA isolation: Preparation of total RNA from Monkeys are often used as experimental animals to livers of Cynomolgus and Japanese monkeys, and assess the metabolism and pharmacokinetics of drugs. reverse transcription (RT) were carried out as described DD exists in four multiple forms (DD1–DD4) in previously.21) Two partial cDNAs for cyDD1 and Japanese monkey liver. DD1 and DD2 have been cyDD4 were ampliêd from the single-strand cDNAs by thought to be the major and minor forms of indanol de- PCR using Pfu DNA polymerase, and forward and hydrogenase, and DD3 and DD4 are identical with alde- reverse primers (Table 1). Because the peptide sequences hyde reductase and 3a-HSD, respectively.25) Asubse- of cyDD1 and cyDD4 were similar to the sequences of quent study on the substrate speciˆcity of DD1 showed human AKR1C1 and AKR1C4, respectively, the for- that the enzyme possesses 3(20)a-HSD activity.26) These ward and reverse primers were designed to anneal the se- ˆndings have suggested that monkey liver DDs cor- quences outside the putative open reading frames of the respond to human 3a-HSD isoenzymes. However, there cDNA species for the two monkey enzymes based on the has been no report of the sequences of the monkey DDs, nucleotide sequences of the corresponding regions of and their properties including substrate speciˆcity, and the cDNAs for the human enzymes. A 936-bp fragment the eŠects of inhibitors and activators have not been ampliêd using a pair of primers, HD1f and HD1r, was well studied compared to those of the human enzymes, conˆrmed to be a partial cDNA for cyDD1 by nucleo- AKR1C1–AKR1C4. Here, we have isolated cDNA spe- tide sequencing.17) The 3?-and5?-ends of the cDNA cies for DD1 and DD4 from livers of both Cynomolgus were generated using the 3?-and5?-RACE kits and and Japanese monkeys, and examined the properties of gene-speciˆc primers (Table 1). Similarly, a cDNA for the recombinant enzymes and the tissue distribution in cyDD4 was generated by RT-PCR using primers (HD4f Japanese monkey. and HD4r), followed by RACE-PCR. The sequences of the coding regions in the cDNAs were re-examined by Materials and Methods sequencing the cDNAs, which had been ampliêd from Materials: Monkey tissues were obtained from the the single-strand cDNA sample obtained with Pfu DNA Primate Research Institute, Kyoto University (Inuyama, polymerase, and primers corresponding to the regions Japan). Lysylendopeptidase (LysC) and isopropyl b-D- near the 3?-and5?-ends of the coding regions (MD1f thiogalactopyranoside were obtained from Wako Pure and MD1r for cyDD1 cDNA, and MD4r and MD4f for Chemicals (Osaka, Japan); Pfu DNA polymerase and cyDD4 cDNA). cDNA species for Japanese monkey Escherichia coli cells were from Stratagene; SuperScript DD1 (jpDD1) and DD4 (jpDD4) were ampliêd from II reverse transcriptase and rapid ampliˆcation of the total RNA sample of the monkey liver by RT-PCR cDNA ends (RACE) kits were from Gibco-BRL; origi- using a set of primers (MD1f and MD1 for DD1 cDNA, nal TA cloning and pCR T7WCT-TOPO TA cloning kits and MD4r and MD4f for DD4 cDNA). were from Invitrogen; and Taq DNA polymerase and Expression and puriˆcation of recombinant DDs: DNA-modifying enzymes were from Takara Shuzou The cDNAs for cyDD1 and jpDD4 were inserted into (Kusatsu, Japan). Pyridine nucleotide coenzymes were pCR T7WTOPO expression plasmids using a pCR T7W purchased from Oriental Yeast Co. (Tokyo, Japan); CT-TOPO TA cloning kit. The constructs were trans- prostaglandin D2 was from Cayman Chemicals (MI, formed into E. coli BL21(DE3)pLysS. The E. coli cells USA); and [9,11,12-3H]5a-pregnan-3a-ol-20-one (65 Ci were cultured in 1 liter of LB medium containing am- Wmmol) was from Daiichi Pure Chemicals (Tokyo, picillin (50 mgWmL) and chloramphenicol (50 mgWmL) at Japan). Medazepam, estazolam, and befunolol were 379C until the turbidity reached 0.4. Then, 1 mM gifts from Nippon Roche Co. (Tokyo, Japan), Takeda isopropyl b-D-thiogalactopyranoside was added and Pharmaceutical Co. (Osaka, Japan), and Kaken Phar- growth of the culture was continued for 6 h. A cell maceutical Co. (Tokyo, Japan), respectively. Other sub- extract was prepared as described previously.27) The pro- strates, inhibitors and activators were obtained from teins in the extract were fractionated by adding

Sigma Chemicals and Fluka Chemie AG. (NH4)2SO4, and the precipitate obtained between 35 and

Protein sequencing: DD1andDD4werepuriˆed 70z (NH4)2SO4 saturation was dissolved in 20 mM from Cynomolgus monkey liver according to the proce- potassium phosphate, pH 7.0, containing 0.2 M KCl, dures for the puriˆcation of the enzymes of Japanese and dialyzed against the buŠer. The enzyme was applied 350 Yu HIGAKI, et al.

Table 1. Nucleotide sequences of forward (f) and reverse (r) primers

Primer Sequence Position

ForPCRforcyDD1cDNA HD1f 5?-CCGAATTCATGGATTCCAAACAGCAGTGT-3? 1–21 HD1r 5?-AGCAAAAATATCAAGGGTCAA-3? 916–936 ForPCRforcyDD4cDNA HD4f 5?-GGGATTTGGCACCTATGC-3? 57–74 HD4r 5?-CATAAGAAAATCCATGACAAC-3? 916–936 For 3?-RACE of cyDD1 cDNA md1f1 5?-GGGATCCCATCGAGAAAAACC-3? 657–677 md1f2 5?-ACGGATCAGAGAGAACATGA-3? 825–844 For 5?-RACE of cyDD1 cDNA md1r1 5?-ACCTCCCATGTGGCAC-3? 434–449 md1r2 5?-CTGGCTTTAGAGACACTGGAAAATG-3? 349–373 md1r3 5?-CATTATTGTAAACATGAGCAGA-3? 151–172 For 3?-RACE of cyDD4 cDNA md4f1 5?-CTGGTTGCCTATAGTGCTCTG-3? 637–657 md4f2 5?-GTCCTGGCCATGAGCTACAA-3? 799–818 For 5?-RACE of cyDD4 cDNA md4r1 5?-TCATCTTTTGGTAGTGGCGT-3? 379–398 md4r2 5?-ACCTCCACAACTCTGTTCC-3? 92–110 ForPCRforjpDD1cDNA MD1f 5?-TTTGCCAGTCAGGCCAGTGA-3? „25–„6 MD1r 5?-CCTCATGCAATGCCCTCCA-3? 975–993 ForPCRforjpDD4cDNA MD4f 5?-TGCATAGCGAAAGAAGTGAC-3? „25–„6 MD4r 5?-TGTCGTGCAACACCCTCTC-3? 975–993 ForexpressionofcyDD1 cyD1Ef 5?-ATGGATTCGAAACATCAGTGTGTG-3? 1–24 cyD1Er 5?-CCGGTACCTTAATATTCATCAGAAAATG-3? 952–972 ForexpressionofjpDD4 jpD1Ef 5?-ATGGATCCCAAATATCAGC-3? 1–19 jpD4Er 5?-CCGGTACCCTAATATTCATCT-3? 959–972 to a Pharmacia Sephadex G-100 column (4.8×80 cm) Enzyme assays: The dehydrogenase and reductase equilibrated with the same buŠer. The enzyme fractions activities of the puriêd enzymes were assayed by meas- were concentrated by ultraˆltration with an Amicon uring the rates of changes in ‰uorescence (at 455 nm YM-10 membrane, dialyzed against 20 mM Tris-HCl, with excitation at 340 nm) and absorbance (at 340 nm), pH8.5,andthenappliedtoanAmiconMatrexRed-A respectively, of NADPH. The standard reaction column (2.0×6 cm) equilibrated with the buŠer. The mixture for dehydrogenase activity consisted of 0.1 M column was washed with the buŠer containing 0.1 M potassium phosphate buŠer, pH 7.4, 0.25 mM NADP+, NaCl, and the enzyme was eluted with the buŠer con- 1.0 mM S-(+)-1,2,3,4-tetrahydronaphth-1-ol (S- taining 0.1 M NaCl and 0.5 mM NADP+.Atthisstep, tetralol) and enzyme, in a total volume of 2.0 mL. The cyDD1 was puriêd to homogeneity, but the jpDD4 activity during the puriˆcation was assayed with 0.1 M preparation contained some other proteins. The jpDD4 glycine-NaOH buŠer, pH 10.0, instead of the phosphate fractions were dialyzed against 5 mM Tris-HCl, pH 8.0, buŠer. The reductase activity for carbonyl compounds and further puriêd by chromatography on a Pharma- was determined with 0.1 M potassium phosphate buŠer, cia Q-Sepharose column (2×20 cm) equilibrated with pH 7.0, and 0.1 mM NADPH as the buŠer and coen- the buŠer. The enzyme was eluted with a linear gradient zyme, respectively. The substrates, inhibitors and of 0–0.1 M NaCl in the buŠer. All the above puriˆca- activators, which were insoluble in water, were ˆrst dis- tion procedures were performed at 49C, and the buŠers solved in methanol, and then added to the reaction mix- were supplemented with 0.5 mM EDTA, 5 mM 2-mer- ture, in which the ˆnal concentration of methanol was captoethanol and 20z (vWv) glycerol to stabilize the en- less than 2.5z. The concentration of methanol did not zymes. aŠect the activity of the monkey and human enzymes. Recombinant AKR1C128) and AKR1C427) were pre- One unit of enzyme activity was deˆned as the amount pared and puriêd as previously described. Protein con- that catalyzed the formation or oxidation of 1 mmol of 29) centrations were determined by Bradford's method NADPH per min at 259C. The apparent Km and Vmax using bovine serum albumin as the standard. values for substrates were determined by Lineweaver- Monkey Dihydrodiol Dehydrogenases 351

Fig. 1. SDS-PAGE of puriˆed DDs. The liver (native) and recombinant (recom.) enzymes (each 2 mg), and molecular mass standards (M) were runona12.5z polyacrylamide slab gel, and then stained with Coomassie brilliant blue R-250.

Burk analysis with the following diŠerent concentra- of the coding regions of cDNA for human AKR1C1 or tions: alicyclic alcohols (0.002–1.0 mM), androstanes AKR1C4. PCR with a pair of primers, HD1f and HD1r, (0.2–40 mM), pregnanes (0.1–10 mM), and non-steroidal ampliêd a 936-bp cDNA fragment, and that with carbonyl compounds (0.5–50 mM). Unless otherwise another pair of primers, HD4f and HD4r, yielded an noted, the kinetic values are the means of at least three 879-bp cDNA fragment. The peptide sequences of determinations. cyDD1 and cyDD4 perfectly matched the regions of the Product identiˆcation: The products of the ox- amino acid sequences deduced from the 936-bp and idoreduction of [3H]-labeled 5a-pregnan-3a-ol-20-one 879-bp cDNA fragments, respectively. Subsequently, and reduction of 5a-pregnane-3,20-dione were identiêd the sequences of the full-length cDNAs for cyDD1 (1225 by TLC analysis as described previously.30) bp) and cyDD4 (1194 bp) were determined by 3?-and5?- RACE with the gene-speciˆc primers. The two cDNA Results species, that contained 972-bp coding regions, were also Peptide sequences of monkey DDs and isolation of ampliêd from the total RNA sample of Japanese their cDNAs: The 105,000×g supernatant of the monkey liver by RT-PCR using primers that annealed homogenate of Cynomolgus monkey liver showed high the 5?-and3?-noncoding regions of the cDNAs for S-tetralol dehydrogenase activity, 0.13 unitWmg, and on cyDD1 and cyDD4. The sequence of the cDNA ampli- the subsequent chromatography on a Q-Sepharose êd using primers MD1f and MD4r showed a diŠerence column the activity was separated into three peaks, of eight nucleotides from that of cDNA for cyDD1, and which correspond to DD1, DD2 and DD4 of Japanese the other cDNA ampliêd with primers MD4f and monkey liver.25) The activity yields for the three peaks, MD4r diŠered from the cDNA for cyDD4 by ˆve designated as cyDD1, cyDD2 and cyDD4, were 67, 3 nucleotides. The proteins encoded by the former and and 7z, and the substrate speciˆcity and inhibitor sen- latter cDNA species of Japanese monkey were designat- sitivity of the DD2 preparation were almost identical to ed as jpDD1 and jpDD4, respectively. those of DD1, as described for DD2 of Japanese The amino acid sequence deduced from the cDNA for monkey liver.25) Therefore, cyDD1 and cyDD4 were fur- cyDD1 was identical to that deduced from the cDNA ther puriêd by consecutive column chromatographic for jpDD1, with the exception of two amino acid substi- fractionation on Matrex Red A and hydroxylapatite. tutions at positions 35 and 216 (Fig. 2). There was only SDS-PAGE analysis of the ˆnal preparations of the two one residue substitution at position 23 between the se- enzymes revealed a single protein band corresponding to quences deduced from the cDNAs for cyDD4 and 36 kDa. (Fig. 1). jpDD4. The sequence identity between cyDD1 and To obtain structural information on cyDD1 and cyDD4 or between jpDD1 and jpDD4 was approximate- cyDD4 for subsequent cDNA cloning, 14 LysC-digested ly 83z. The monkey enzymes showed high sequence peptides and two BrCN-cleaved peptides of the respec- identity with human AKR1C1–AKR1C4 among the tive enzymes were isolated and sequenced. The peptide AKR family proteins. The sequence identities of the sequences of cyDD1 and cyDD4 were quite similar to monkey enzymes with AKR1C1, AKR1C2, AKR1C3 the regions of the amino acid sequences of human and AKR1C4 were 94, 93, 87 and 83z, respectively, for AKR1C1 and AKR1C4, respectively (Fig. 2). Therefore, cyDD1 or jpDD1, the respective values being 83, 81, 84 we initially ampliêd cDNA fragments from the total and 94z for cyDD4 or jpDD4. RNA sample of Cynomolgus monkey liver by RT-PCR Puriˆcation and properties of recombinant enzymes: using primers that annealed the 5?-and3?-end regions Because the amino acid sequences of cyDD1 and cyDD4 352 Yu HIGAKI, et al.

Fig. 2. Alignment of the deduced amino acid sequences of monkey DDs with human AKR1C1–AKR1C4. Identical amino acid residues between cyDD1 and the other proteins are denoted by hyphens. The underlined regions of the sequences completely matched the sequences of peptides derived from Lys-C digestion (straight line) and BrCN cleavage (wavy line) of cyDD1 and cyDD4 of Cynomolgus monkey liver. The two residues at positions 35 and 216 of the cyDD1 sequence are replaced by Ile and Tyr, respectively, in the jpDD1 sequence, and only one residue replacement of Thr to Ser at position 23 is observed in the sequences of cyDD4 and jpDD4. The nucleotide sequences have been deposited in DDBJ databases under the following accession numbers: AB070209 (cyDD1), AB070211 (cyDD4), AB070210 (jpDD1), and AB070212 (jpDD4). were almost identical to those of the respective enzymes tivity toward progesterone, befunolol or prostaglandin of Japanese monkey, recombinant cyDD1 and jpDD4 D2, and reduced both 3-ketoandrostanes and 3- were expressed from the respective cDNAs in E. coli ketopregnanes more highly than 20-ketopregnanes. Be- cells and puriˆed to elucidate the diŠerence in properties cause cyDD1 and jpDD4 exhibited high sequence identi- between the two types of enzyme. The respective yields ty with human AKR1C1 and AKR1C4, respectively, the (amounts) and speciˆc activities of the ˆnal prepara- kinetic constants of the monkey enzymes were com- tions puriˆed from one liter of cultured cells were 35z pared with those of the human enzymes that were deter- (10 mg) and 3.9 unitsWmg for recombinant cyDD1, and mined under the same conditions. The substrate 38z (6 mg) and 3.3 unitsWmg for recombinant jpDD4. speciˆcities of the two monkey enzymes were essentially The enzyme preparations gave single bands on SDS- identical to those of the respective human counterparts,

PAGE, their molecular masses (36 kDa) being the same except that jpDD4 showed higher kcat values for most of as those of cyDD1 and cyDD4 puriˆed from the liver the substrates than AKR1C4 did. It should be noted that (Fig. 1). jpDD4 and AKR1C4, but not cyDD1 and AKR1C1, The two monkey enzymes oxidized both 3a-and20a- showed high substrate inhibition by the 20a-hydrox- hydroxysteroids, but diŠered from each other in the ypregnanes and 20-ketopregnanes at concentrations of kinetic constants (Table 2). While cyDD1 e‹ciently above 3–7 times the respective Km values. oxidized 20a-hydroxypregnanes among the steroid To conˆrm the bifunctional activities of the monkey substrates, jpDD4 exhibited high kcat WKm values for 3a- enzymes, the products of the enzymatic oxidation and hydroxypregnanes as well as for 3a-hydroxyan- reduction of representative substrates were analyzed by drostanes. cyDD1 accepted S-indan-1-ol, S-tetralol and TLC. When cyDD1 and jpDD4 were incubated with 4-chromanol as good non-steroidal substrates, whereas [3 H]-labeled 5a-pregnan-3a-ol-20-one, 5a-pregnane- jpDD4 showed low kcat WKm values for the non-steroidal 3,20-dione and 5a-pregnane-3a,20a-diol were detected substrates, except for R-tetralol and trans-benzene as the products of oxidation and reduction, respectively. dihydrodiol. cyDD1 and AKR1C1 also exhibited low On the reduction of 5a-pregnane-3,20-dione by cyDD1 dehydrogenase activities toward 5a-androstan-17b-ol-3- and jpDD4, both 20a-hydroxy and 3a-hydroxy metabo- „1 one (apparent kcat values of 0.05 and 0.04 min , respec- lites were initially produced by cyDD1 and jpDD4, tively), but the 17b-HSD activity was not detected for respectively, and then 5a-pregnane-3a,20a-diol was jpDD4 or AKR1C4. In the reverse reaction, cyDD1 formed on prolonged incubation, as observed with reduced all the steroidal and non-steroidal substrates AKR1C1 and AKR1C4.30) listed in Table 3, and showed higher kcat WKm values for EŠects of inhibitors and activators: Phenolphtha- 3- and 20-ketopregnanes than for 3-ketoandrostanes. In lein, 1,10-phenanthroline and the drugs listed in Table 4 contrast, jpDD4 did not exhibit signiˆcant reductase ac- have been reported to inhibit human AKR1C1 and Monkey Dihydrodiol Dehydrogenases 353

Table 2. Kinetic constants in the oxidation of hydroxysteroids and non-steroidal alcohols by monkey DDs, AKR1C1 and AKR1C4a)

cyDD1 (AKR1C1) jpDD4 (AKR1C4) Substrate Km kcat kcat WKm Km kcat kcat WKm ( mM) (min„1)(min„1mM„1)(mM) (min„1)(min„1mM„1)

3a-Hydroxyandrostanes 5a-Androstan-3a-ol-17-one 10 (15) 0.39 (0.21) 0.04 (0.01) 0.7 (0.5) 12 (2.6) 17 (5.2) 5a-Androstane-3a,17b-diol 13 (18) 0.5 (0.3) 0.04 (0.02) 0.6 (0.8) 13 (5.2) 21 (6.2) 3a- or 20a-Hydroxypregnanes 5a-Pregnan-3a-ol-20-one 3.7 (3.3) 0.22 (0.08) 0.06 (0.02) 0.3 (0.3) 9.2 (4.0) 31 (13) 5b-Pregnan-3a-ol-20-one 2.8 (4.5) 0.40 (0.38) 0.14 (0.09) 0.3 (0.4) 4.9 (1.6) 16 (4.0) 5a-Pregnane-3a,21-diol-20-one 2.8 (2.6) 0.19 (0.22) 0.07 (0.09) 0.6 (0.3) 23 (4.2) 38 (14) 5a-Pregnan-20a-ol-3-one 2.1 (2.7) 8.6 (9.1) 4.1 (3.4) 0.2 (0.2) 1.5 (0.64) 7.5 (3.2) 5b-Pregnan-20a-ol-3-one 1.5 (1.0) 6.3 (5.6) 4.2 (5.6) 0.2 (0.3) 0.47 (0.60) 2.4 (2.0) 4-Pregnen-20a-ol-3-one 12 (16) 4.3 (9.1) 0.35 (0.56) 0.3 (0.3) 0.76 (0.34) 2.5 (1.1) Non-steroidal alcohols S-Indan-1-ol 76 (50) 34 (25) 0.45 (0.50) 420 (146) 20 (6.1) 0.05 (0.04) S-Tetralol 29 (7.4) 32 (26) 1.1 (3.7) 307 (105) 36 (9.7) 0.12 (0.09) R-Tetralol 56 (180) 1.4 (1.1) 0.02 (0.01) 382 (160) 42 (11) 0.08 (0.07) 4-Chromanol 73 (104) 48 (39) 0.65 (0.38) 340 (650) 2.5 (6.2) 0.007 (0.10) trans-Benzene dihydrodiol 720 (820) 4.1 (8.7) 0.006(0.01) 25 (22) 1.5 (0.6) 0.06 (0.03) a)The dehydrogenase activity was assayed at pH 7.4, and the kinetic values of the human enzymes are shown in parentheses.

Table 3. Comparison of the kinetic constants for ketosteroids among monkey DDs, AKR1C1 and AKR1C4a)

cyDD1 (AKR1C1) jpDD4 (AKR1C4) Substrate Km kcat kcat WKm Km kcat kcat WKm ( mM) (min„1)(min„1mM„1)(mM) (min„1)(min„1mM„1)

3-Ketoandrostanes 5a-Androstan-17b-ol-3-one 11 (12) 1.1 (1.1) 0.10(0.09) 0.8 (0.7) 2.4 (1.4) 3.0 (2.0) 5a-Androstane-3,17-dione 2.2 (3.3) 1.4 (1.6) 0.64(0.48) 1.3 (1.4) 2.7 (1.8) 2.1 (1.3) 3- andWor 20-Ketopregnanes 5a-Pregnan-20a-ol-3-one 0.9 (0.7) 0.90 (0.60) 1.1 (0.86) 0.8 (1.1) 2.9 (1.1) 3.6 (1.0) 5b-Pregnan-20a-ol-3-one 1.7 (1.3) 3.2 (2.2) 1.9 (1.7) 0.7 (1.2) 3.2 (1.4) 4.6 (1.2) 5a-Pregnan-3a-ol-20-one 2.3 (2.0) 5.0 (4.4) 2.2 (2.2) 0.4 (0.9) 0.43(0.60) 1.1 (0.67) 5b-Pregnan-3a-ol-20-one 3.8 (3.3) 9.8 (11) 2.6 (3.3) 0.4 (0.5) 0.78(0.73) 1.8 (1.5) 5a-Pregnane-3,20-dione 0.5 (1.1) 2.8 (3.0) 5.6 (2.7) 0.9 (0.6) 4.8 (1.2) 5.3 (2.0) 5a-Pregnane-21-ol-3,20-dione 1.9 (1.2) 0.92 (0.50) 0.48(0.42) 0.8 (1.7) 4.2 (2.1) 5.3 (1.2) Progesterone 0.7 (1.3) 1.3 (1.5) 1.8 (1.2) na (na)b) Non-steroidal compounds a-Tetralone 1.0 (1.3) 3.4 (3.0) 3.4 (2.3) 38 (35) 1.5 (1.7) 0.04(0.05) 4-Chromanone 3.1 (3.1) 6.7 (7.2) 2.2 (2.3) 0.73(0.48)c) Befunolol 27 (41) 5.6 (6.0) 0.21(0.15) na (na)

Prostaglandin D2 11 (12) 1.1 (1.0) 0.1 (0.08) na (na) a)The activity was assayed at pH 7.0, and the kinetic values of the human enzymes are shown in parentheses. b)na, no signiˆcant activity was detected with 10 mM substrate. c) The kcat values were calculated with the activities for 10 mM substrate.

AKR1C4.10,13,27,28) These inhibitors also inhibited cyDD1 AKR1C4 has been reported to be activated by sul- and jpDD4. When other compounds tested as inhibitors fobromophthalein (BSP), cloˆbric acid derivatives28) of the enzymes, 3?,3!,5?,5!-tetrabromophenolphthalein and thyroxines.31) The activators of AKR1C4 enhanced (TBPP) and benzbromarone were found to be more po- the dehydrogenase activity of jpDD4, but not that of tent inhibitors of AKR1C1 and cyDD1 than the known cyDD1. inhibitors. In addition, o-cresolphthalein was a potent Tissue distribution: The distribution of DD1 and inhibitor of AKR1C4 and jpDD4. No signiˆcant diŠer- DD4 in Japanese monkey tissues was assessed by RT- ence in the IC50 (concentration required for 50z inhibi- PCR (Fig. 3). The expression of the jpDD1 transcript tion) values for all the inhibitors was observed between was detected in liver, adrenal gland, intestine and kid- cyDD1 and AKR1C1, or between jpDD4 and AKR1C4. ney, whereas jpDD4 mRNA was detected only in liver 354 Yu HIGAKI, et al.

Fig. 3. Expression of mRNAs for DD1 and DD4 in Japanese monkey tissues. RT-PCR with MD1f and MD1r (a), jpD1Ef and jpD4Er (b), or the speciˆc primers for b-actin cDNA (c) was performed for the total RNA from the monkey tissues. The PCR products were subjected to electropho- resis on 2z (wWv) agarose gels and visualized by staining with ethidium bromide. Lanes: 1, cerebrum; 2, cerebellum; 3, lung; 4, liver; 5, intestine; 6, adrenal gland; 7, kidney.

Table 4. EŠects of inhibitors and activators on the dehydrogenase DD1 and DD4 of monkey liver indicate that the two activities of monkey DDs, AKR1C1 and AKR1C4 enzymes belong to the AKR family and act as 3(20)a-

a) HSDs.DD1andDD4showthehighestaminoacidse- IC50 value ( mM) Inhibitor quence identity with AKR1C1 and AKR1C4, respective- cyDD1 (AKR1C1) jpDD4 (AKR1C4) ly, among the AKR family proteins. In addition, the TBPP 0.03 (0.03) 0.5 (1.0) substrate speciˆcity and inhibitor sensitivity of cyDD1 o-Cresolphthalein 0.3 (0.1) 0.04 (0.06) and jpDD4 were similar to those of AKR1C1 and Phenolphthalein 4.8 (4.6) 0.09 (0.08) AKR1C4, respectively. Thus, DD1 and DD4 are homo- 1,10-Phenanthroline 45 (65) 810 (1500) logues of human AKR1C1 and AKR1C4, respectively. Benzbromarone 0.03 (0.05) 0.4 (0.7) Medroxyprogesterone 1.0 (0.7) 0.05 (0.03) Although DD1 was previously classiêd as indanol de- Flufenamic acid 5.5 (5.3) 51 (110) hydrogenase in the Enzyme Nomenclature system, we Medazepam 7.0 (4.7) n.i (n.i)b) propose that the enzyme should be named 3(20)a-HSD Hexestrol 14 (8.3) 0.7 (0.6) because of its high catalytic e‹ciency for steroids with Estazolam 22 (6.6) n.i (n.i) 3(20)a-hydroxy and 3(20)-keto groups. The kinetic con- Dexamethasone n.i (n.i) 0.9 (1.6) Betamethasone n.i (n.i) 2.1 (2.4) stants of cyDD1 and AKR1C1 for these substrates were almost identical, suggesting that the 19 residues diŠerent Stimulation percentagec) Activator between the cyDD1 and AKR1C1 sequences are not cyDD1 (AKR1C1) jpDD4 (AKR1C4) responsible for substrate binding of the enzymes. On the other hand, the k valuesformostofthesubstratesof 10 mM D-Thyroxine 0 (0) 183 (318) cat 10 mM BSP 0 (0) 182 (310) jpDD4 were higher than those of AKR1C4, while the Km 0.4 mM Cloˆbric acid 0 (0) 53 (140) values of the two enzymes were similar. This would sug- a) gest that some of the 18 residues that diŠer between the The IC50 values of the human enzymes, shown in parentheses, are taken from previous papers,10,13,27,28,30) except that the values for monkey and human enzymes are important for proper TBPP, o-cresolphthalein and benzbromarone were determined in orientation of substrates to achieve high catalytic tur- this study. nover with a minimal eŠect on substrate a‹nity. b)n.i, no signiˆcant inhibition (less than 30z)wasobservedwith In humans, two further 3a-HSDs, AKR1C2 and 0.1 mM benzodiazepines and 50 mM steroidal anti-in‰ammatory AKR1C3, have been identiêd.14–16) Recently, a cDNA drugs. c) for monkey type 5 17b-HSD, which exhibits high amino (v„vo) Wvo×100, where v and vo represent the velocities in the presence and absence, respectively, of the activator. acid sequence identity with AKR1C3, has been cloned,33) although it remains unknown whether or not the recombinant enzyme possesses the activities of 3a- and kidney. We also examined whether or not an HSD and prostaglandin F synthase,21,22) which are mRNA species for another enzyme similar to AKR1C2 characteristics of human AKR1C3. This ˆnding, is expressed in Cynomolgus and Japanese monkey livers together with the present identiˆcation of monkey by RT-PCR using the primers listed in Table 1 and other homologues of AKR1C1 and AKR1C4, suggests the primers that anneal regions of AKR1C2 cDNA.32) The possibility that a homologue of AKR1C2 is also ex- DNA sequencing showed that the ampliêd cDNAs are pressed in monkey tissues. The amino acid sequences of the same as those for cyDD1 and jpDD1. AKR1C1 and AKR1C2 are extremely similar (they diŠer only by seven residues),13,14) and mRNA for AKR1C2 is Discussion ubiquitously expressed in many human tissues.18,32) Con- The present cloning and expression of two cDNAs for sidering the high nucleotide sequence identity between Monkey Dihydrodiol Dehydrogenases 355

35) DD1 and AKR1C1, and between DD4 and AKR1C4, it teers. Because the IC50 value of the drug for AKR1C1 should be possible to amplify a cDNA for a monkey is much lower than its plasma concentration, long-term homologue of AKR1C2 by RT-PCR with the primers therapy with this drug may modulate the metabolism of for cyDD1 or other primers that anneal regions of steroids and xenobiotics mediated by this enzyme. AKR1C2 cDNA.32) However, no cDNA fragment distinct from that for DD1 was ampliˆed from the total Acknowledgement: This work was supported by a RNA samples of Cynomolgus and Japanese monkey Grant-in-Aid for Scientiˆc Research (B) from the Minis- livers by RT-PCR with these primers. The minor form try of Education, Science, Sports and Culture, Japan. (DD2) in Japanese monkey liver seems to be a candidate References for monkey homologue of AKR1C2, but it has been suggested to be a product of post-translational modiˆ- 1) van Heyningen, R.: Naphthalene cataract in rats and cation of DD1.25) In addition, the substrate speciˆcity rabbits: A resumáe. Exp. Eye Res., 28: 435–439 (1979). and inhibitor sensitivity of the partially puriˆed DD2 2) Penning, T. M., Burczynski, M. E., Hung, C.-F., preparation of Cynomolgus monkey liver were almost McCoull, K. D., Palackal, N. T. and Tsuruda, L. S.: identical to those of cyDD1. The similarity in the Dihydrodiol dehydrogenases and polycyclic aromatic properties of DD1 and DD2 is in contrast to the reports hydrocarbon activation: Generation of reactive and redox active o-quinones. Chem. Res. Toxicol., 12: 1–18 that AKR1C1 shows steroid speciˆcity and inhibitor (1999). 13,14,18,30) sensitivity distinct from those of AKR1C2. 3) W äorner, W. and Oesch, F.: Identity of dihydrodiol de- Therefore, a homologue of AKR1C2 may not exist at hydrogenase and 3a-hydroxysteroid dehydrogenase in least in the livers of the two strains of monkey, although rat but not in rabbit liver cytosol. FEBS Lett., 170: we cannot rule out its expression in other tissues. 263–267 (1984). DD1 was the major form of DD in Cynomolgus 4) Penning, T. M., Mukharji, I., Barrows, S. and Talalay, monkey liver as reported on puriˆcation of this enzyme P.: Puriˆcation and properties of a 3a-hydroxysteroid form as DD25) and indanol dehydrogenase26) from dehydrogenase of rat liver cytosol and its inhibition by Japanese monkey liver. The high expression of DD1, anti-in‰ammatory drugs. Biochem. J., 222: 601–611 the homologue of AKR1C1, in monkey liver is an addi- (1984). 5) Sawada, H., Hara, A., Nakagawa, M., Tsukada, F., tional diŠerence from in the case of human liver, in Ohmura, M. and Matsuura, K.: Separation and proper- which at least AKR1C1, AKR1C2 and AKR1C4 are ex- ties of multiple forms of dihydrodiol dehydrogenase 18) 10,11) pressed almost equally at the mRNA and activity from hamster liver. Int. J. Biochem., 21: 367–375 (1989). levels. In addition, the distribution pattern of the 6) Sawada, H., Hara, A., Nakayama, T., Nakagawa, M., jpDD1 transcript in Japanese monkey tissues diŠers Inoue, Y., Hasebe, K. and Zhang, Y.-P.: Mouse liver di- from the ubiquitous expression of the AKR1C1 tran- hydrodiol dehydrogenases. Identity of the predominant script in human tissues.18,34) Furthermore, the detection andaminorformwith17b-hydroxysteroid de- of the jpDD4 transcript in both liver and kidney is hydrogenase and aldehyde reductase. Biochem. Phar- diŠerent from the liver-speciˆc expression of AKR1C4. macol., 37: 453–458 (1988). Thus, the diŠerences in tissue distribution of the two 7) Hara, A., Hasebe, K., Hayashibara, M., Matsuura, K., types of enzyme between humans and Japanese Nakayama, T. and Sawada, H.: Dihydrodiol dehydrogenases in guinea pig liver. Biochem. Pharmacol., monkeys suggest that their contributions in each tissue 35: 4005–4012 (1986). to the metabolism of steroids and xenobiotics diŠer be- 8) Matsuura, K., Hara, A., Nakayama, T., Nakagawa, M. tween the two primates. and Sawada, H.: Puriˆcation and properties of two mul- Finally, it is interesting to note that benzbromarone tiple forms of dihydrodiol dehydrogenase from guinea- and TBPP are potent inhibitors of both AKR1C1 and pig testis. Biochim. Biophys. Acta, 912: 270–277 (1987). monkey DD1. The inhibitory selectivity for the enzymes 9) Aoki, S., Ishikura, S., Asada, Y., Usami, N. and Hara, is further supported by our preliminary ˆnding that the A.: Identity of dimeric dihydrodiol dehydrogenase as two compounds weakly inhibit the activities of NADP+-dependent D-xylose dehydrogenase in pig liver. AKR1C2, AKR1C3 and several other members of the Chemico-Biol. Res. Interact., 130–132: 775–784 (2001). AKR family (Y. Higaki and N. Usami, unpublished 10) Hara, A., Taniguchi, H., Nakayama, T. and Sawada, results). Thus, the two compounds may be useful inhibi- H.: Puriˆcation and properties of multiple forms of di- tors for discriminating AKR1C1 and DD1 from the hydrodiol dehydrogenase from human liver. J. Biochem. (Tokyo), 108: 250–254 (1990). other isoenzymes or multiple forms of 3 HSD with DD a 11) Deyashiki, Y., Taniguchi, H., Amano, T., Nakayama, activity in human and monkey tissues. In addition, ben- T., Hara, A. and Sawada, H.: Structural and functional zbromarone is clinically used for the treatment of comparison of two human liver dihydrodiol de- hyperuricaemia, and its maximum plasma concentra- hydrogenases associated with 3a-hydroxysteroid de- tion was reported to be 4.8 mM after the administration hydrogenase activity. Biochem. J., 282: 741–746 (1992). of a single oral dose (100 mg) to seven healthy volun- 12) Jez, J. M. and Penning, T. M.: The aldo-keto reductase 356 Yu HIGAKI, et al.

(AKR) superfamily: an update. Chemico-Biol. Res. In- dehydrogenases, carbonyl reductase and aldehyde reduc- teract., 130–132: 499–525 (2001). tase of human liver. Biochem. Pharmacol., 50: 221–227 13) Hara, A., Matsuura, K., Tamada, Y., Sato, K., Miyabe, (1995). Y., Deyashiki, Y. and Ishida, N.: Relationship of human 24) Breyer-PfaŠ, U. and Nill, K.: High-a‹nity stereoselec- liver dihydrodiol dehydrogenases to hepatic bile-acid- tive reduction of the enantiomers of ketotifen and keton- binding protein and an oxidoreductase of human colon ic nortriptyline metabolites by aldo-keto reductases from cells. Biochem. J., 313: 373–376 (1996). human liver. Biochem. Pharmacol., 59: 249–260 (2000). 14) Dufort, I., Soucy, P., Labrie, F. and Luu-The, V.: 25) Nakagawa,M.,Harada,T.,Hara,A.,Nakayama,T. Molecular cloning of human type 3 3a-hydroxysteroid and Sawada, H.: Puriˆcation and properties of multiple dehydrogenase that diŠers from 20a-hydroxysteroid de- forms of dihydrodiol dehydrogenase from monkey liver. hydrogenase by seven amino acids. Biochem. Biophys. Chem. Pham. Bull., 37: 2852–2854 (1989). Res. Commun., 228: 474–479 (1996). 26) Hara, A., Nakagawa, M., Taniguchi, H. and Sawada, 15) Khanna, M., Qin, K.-N., Wang, R. W. and Cheng, H.: 3(20)a-Hydroxysteroid dehydrogenase activity of K.-C.: Substrate speciˆcity, gene structure, and tissue- monkey liver indanol dehydrogenase. J. Biochem. speciˆc distribution of multiple human 3a-hydrox- (Tokyo), 106: 900–903 (1989). ysteroid dehydrogenases. J. Biol. Chem., 270: 20162– 27) Deyashiki, Y., Tamada, Y., Miyabe, Y., Nakanishi, M., 20168 (1995). Matsuura, K. and Hara, A.: Expression and kinetic 16) Lin, H.-K., Jez, J. M., Schlegel, B. P., Peehl, D. M., properties of a recombinant 3a-hydroxysteroidWdi- Pachter, J. A. and Penning, T. M.: Expression and hydrodiol dehydrogenase isoenzyme of human liver. J. characterization of recombinant type 2 3a-hydrox- Biochem. (Tokyo), 118: 285–290 (1995). ysteroid dehydrogenase (HSD) from human prostate: 28) Matsuura, K., Hara, A., Deyashiki, Y., Iwasa, H., Demonstration of bifunctional 3aW17b-HSD activity and Kume, T., Ishikura, S., Shiraishi, H. and Katagiri, Y.: cellular distribution. Mol. Endocrinol., 11: 1971–1984 RolesoftheC-terminaldomainsofhumandihydrodiol (1997). dehydrogenase isoforms in the binding of substrates and 17) Deyashiki, Y., Ogasawara, A., Nakayama, T., modulators: Probing with chimaeric enzymes. Biochem. Nakanishi, M., Miyabe, Y., Sato, K. and Hara, A.: J., 336, 429–436 (1998). Molecular cloning of two human liver 3a-hydroxysteroid 29) Bradford, M. M.: A rapid and sensitive method for the Wdihydrodiol dehydrogenase isoenzymes that are identi- quantitation of microgram quantities of protein utilizing cal with chlordecone reductase and bile acid binder. the principle of protein-dye binding. Anal. Biochem., Biochem. J., 299: 545–552 (1994). 72: 248–254 (1976). 18) Penning, T. M., Burczynski, M. E., Jez, J. M., Hung, 30) Usami, N., Yamamoto, T., Shintani, S., Ishikura, S., C.-F., Lin, H.-K., Ma, H., Moode, M., Palackal, Higaki, Y., Katgiri, Y. and Hara, A.: Substrate speciˆci- N. and Patnam, K.: Human 3a-hydroxysteroid de- ty of human 3(20)a-hydroxysteroid dehydrogenase for hydrogenase isoforms (AKR1C1–AKR1C4) of the aldo- neurosteroids and its inhibition by benzodiazepines. keto reductase superfamily: functional plasticity and Biol. Pharm. Bull., 25: 441–445 (2002). tissue distribution reveal roles in the inactivation and 31) Yamamoto, T., Nozaki, A., Shintani, S., Ishikura, S., formation of male and female sex hormones. Biochem. Katagiri, Y. and Hara, A.: Structure-speciˆc eŠects of J., 351: 67–77 (2000). thyroxine analogs on human liver 3a-hydroxysteroid de- 19) Zhang,Y.,Dufort,I.,Rheault,P.andLuu-The,V.: hydrogenase. J. Biochem. (Tokyo), 128: 121–128 (2000). Characterization of a human 20a-hydroxysteroid de- 32) Shiraishi, H., Ishikura, S., Matsuura, K., Deyashiki, Y., hydrogenase. J. Mol. Endocrinol., 25: 221–228 (2000). Ninomiya,M.,Sakai,S.andHara,A.:Sequenceofthe 20) Dufort, I., Rheault, P., Huang, X.-F., Soucy, P. and cDNA of a human dihydrodiol dehydrogenase isoform Luu-The, V.: Characteristics of a highly labile human (AKR1C2) and tissue distribution of its mRNA. type 5 17b-hydroxysteroid dehydrogenase. Endocrinolo- Biochem. J., 334: 399–405 (1998). gy, 140: 568–574 (1999). 33) Luu-The, V., Dufort, I., Pelletier, G. and Labrie, F.: 21) Matsuura, K., Shiraishi, H., Hara, A., Sato, K., Type 5 17b-hydroxysteroid dehydrogenase: Its role in the Deyashiki, Y., Ninomiya, M. and Sakai, S.: Identiˆca- formation of androgens in women. Mol. Cell. En- tion of a principal mRNA species for human 3a-hydrox- docrinol., 171: 77–82 (2001). ysteroid dehydrogenase isoform (AKR1C3) that exhibits 34) Shiraishi, H., Matsuura, K., Kume, T. and Hara, A.:

high prostaglandin D2 11-ketoreductase activity. J. Expression of mRNAs for dihydrodiol dehydrogenase Biochem. (Tokyo), 124: 940–946 (1998). isoforms in human tissues. In Weiner, H., Maser, E., 22) Suzuki-Yamamoto, T., Nishizawa, M., Fukui, M., Crabb, D. W. and Lindahl, R. (eds.): Enzymology and Okuda-Ashitaka, E., Nakajima, T., Ito, S. and Molecular Biology of Carbonyl Metabolism 7,New Watanabe, K.: cDNA cloning, expression and charac- York, Kluwer AcademicWPlenum Publishers, 1999, pp. terization of human prostaglandin F synthase. FEBS 539–544. Lett., 462: 335–340 (1999). 35)Ferber,H.,Vergin,H.andHitzenberger,G.:Phar- 23) Ohara, H., Miyabe, Y., Deyashiki, Y., Matsuura, K. macokinetics and biotransformation of benzbromarone and Hara, A.: Reduction of drug ketones by dihydrodiol in man. Eur. J. Clin. Pharmacol., 19: 431–435 (1981).